Multi-depot battery electric bus scheduling and charging coordination under resource limitations

Battery electric buses (BEBs) have gained significant popularity in metropolitan cities due to their environmental benefits. However, their limited range and long charging times pose challenges in optimizing vehicle scheduling and charging plans. To address these challenges, this study proposes a joint optimization model for BEB scheduling and charging across multiple lines and depots, incorporating charging infrastructure capacity constraints. The model employs a time-space network representation while innovatively eliminating vehicle-indexed variables, yet still accurately tracks state-of-charge (SOC) dynamics. We develop an adaptive large neighborhood search (ALNS) algorithm enhanced with two key sub-routines: (1) an SOC adjustment mechanism during the repair phase and (2) a charger/power allocation adjustment procedure. These sub-routines enable dynamic coordination between charging and scheduling decisions throughout the iterative optimization process. The proposed framework is validated using real-world operational data from Jiading District, Shanghai, China. Computational experiments demonstrate that our ALNS algorithm achieves an 88.7 % reduction in solution time compared to GUROBI for a 105-trip instance while maintaining solution quality. Moreover, the method scales effectively, solving a large-scale 460-trip scenario within 0.6 h.